The detection of airborne water droplets and their classification according to droplet size is an important function of an in-flight icing conditions detector. Current ice protection devices on aircraft, such as inflatable boots, are well-suited for ice accumulation from small droplets (e.g. <40 μm mean value diameter) but may not provide protection from ice accumulation when the impinging droplets are large. In particular, the ability to discriminate supercooled large droplets (SLD) is quickly becoming recognized as a critical safety feature for an icing conditions sensor. SLD are typically greater than 40 μm diameter and are well below the freezing temperature of water. When they strike the leading edge of an airplane wing, they tend to roll beyond the leading end and freeze in locations inaccessible to anti-icing devices but critical to the control of the aircraft. Supercooled large droplets are believed to have caused some aircraft accidents, such as the fatal crash of an ATR-72 in Roselawn, Ind. in 1994.
Soft targets with a high density of scattering sites (such as clouds) will produce multiple scattering when they are probed by a laser beam. For multiple scattering, light rays experience two or more scattering events before returning to the lidar receiver. Most analyses of lidar multiple scattering assume that each detected ray experiences numerous small-angle forward scatterings (both while propagating away from and towards the lidar) and one single large-angle (˜180°) scattering event that is responsible for its backscatter towards the lidar receiver. The small-angle forward scatterings are due primarily to diffraction of the light around the particles, and these small angles are largely responsible for the increased field-of-view of the received light as the laser beam penetrates the soft target. In the process of multiple scattering, the rays diffuse laterally, and the received field-of-view will expand beyond the laser divergence, depending on the size distribution and density of the scattering particles that comprise the soft target.
The general relationship between the particle diameter (d), the laser wavelength (λ), and the forward-scattering diffraction angle (β) is:β∝(λ/d)
This is a simple proportional relationship between droplet diameter and scattering angle. Within a cloud, however, there is a distribution of water droplet sizes, and the scattering angles will vary according to this distribution. Generally speaking, however, small particles produce large scattering angles, and vice versa.
FIG. 1 presents a simplified view of the field-of-view as a lidar beam penetrates distance x into a cloud 50 located a distance R from the receiver 52. If the scattering angle is β, then the field-of-view θ can be obtained from:tan(θ)=x tan(β)/(R+x)≈xλ(R+x)d, in the limit of small θ and β
For the case that R=1000 m, x=200 m, λ=1 μm, and d=5 μm (typical of a water cloud), the field of view θ is approximately 40 mrad, which corresponds to the maximum field-of-view employed by prior art multiple field of view lidar systems. However, for supercooled large droplets, droplet sizes range from 50 μm to over 100 μm. In a cloud of 40 μm droplets, the field-of-view decreases to 5 mrad; for 100 μm and larger, it is less than 2 mrad. The inverse relationship of field-of-view with droplet size means that the multiple fields-of-view generated by large droplets crowd close together near the single-scattering field-of-view generated naturally by the divergence of the laser beam.
FIG. 2 shows how the multiple fields-of-view generated by droplets reflecting backscattered light appears at the focal plane. An outgoing collimated light beam 54 illuminates the droplets and the backscattered light 56 from the droplets passes through one or more receiver lenses 58 after which it is received by in a detector region, generally shown as 60, arranged along the optical axis A. At the upper half of the detector's focal plane 62, multiple fields of view map into concentric rings, generally shown as 64.
The concept behind a multiple field-of-view (MFOV) detector is to place multiple detector elements into the focal plane of the receiver optic and simultaneously measure the backscatter from the various fields of view. In the focal plane, the various FOVs occupy different spatial locations, with the distance from the optical axis (y) being proportional to the FOV according to the relation:y=fθwhere f is the focal length of the receiver optic. For a lidar with a 2″ diameter, f/2.5 receiver lens, the displacement is 63 μm for every 0.5 mrad angle with regard to the optical axis of the lidar.
U.S. Pat. No. 5,239,352 (Bissonnette) discloses a prior art receiver for detecting MFOV lidar backscatter. FIGS. 3 and 4 show that this prior art receiver 71 has a multi-element radiation detector 73 located in the focal plane “f” of the receiving optics 72 having optical axis 74. The detector 73 consists of a number of concentric circular silicon detector elements (PIN photodiodes) 73-1, 73-2, 73-3 and 73-4. As a result of the four separate detector elements, the receiver 71 can differentiate received backscattered radiation signals between several fields of view. A backscattered signal received for any field of view larger than the divergence of the lidar's laser beam is due to multiple scattering.
The bandwidth of the detector elements is sufficiently high to ensure range resolution of <5 meters as the beam penetrates the cloud. In this detector, each detector element integrates the signal over a given range of field-of-view and generates a single value. The four concentric detector elements cover the following fields-of-view:                73-1 0-3.75 mrad        73-2 3.75-12.5 mrad        73-3 12.5-25.0 mrad        73-4 25.0-37.5 mrad        
Detector element 73-1 measures the entire single scattering signal with some multiple scattering as well; detector elements 73-2 through 73-4 measure only the multiple scattering. However, for the detection of supercooled large droplets, the fixed FOV at 3.75 mrad might be a limitation since most of the useful scattering information may be completely contained within this single FOV, which also contains the entire single scattering signal. Thus, there is no way to distinguish multiple scattering due to large droplets from that due to single scattering. In addition, the FOVs are fixed and cannot be reconfigured.
U.S. Pat. No. 4,893,003 (Hays) discloses a circle-to-line interferometer optical system (CLIO) for use with a Fabry-Perot interferometer. As seen in FIGS. 5-6, a CLIO system includes a conical reflector segment 80 that is provided with an interior conical reflective surface 81. The conical reflective surface 81 is oriented so as to reflect incoming parallel light rays 83 produced by a Fabry-Perot Interferometer and containing circular fringe information 82. The light rays 83 propagate in a direction substantially parallel to a conical axis 84 of the conical reflector segment 80. The circular fringe information 82 is converted into linear information when the reflected light rays 83 are received by a conventional linear array detector 87, such as a charge coupled device of the sort used in spectroscopic analysis. The radii of the interferometer fringes depend on the spacing of the interferometer's reflective surfaces, the speed of the particles that reflect light into the spectrometer, the wavelength of the light, and on the phase coherence of the light entering the interferometer. The apex of the cone may be situated where the conical axis 84 intersects the focal plane of the circular fringe pattern 82. The azimuthal angle of the detected circular fringe pattern 82 may be reduced with the use of a tele-kaleidoscope 86 (FIG. 6) comprising a predetermined arrangement of mirrors 85. A right-angle cone with a reflective surface 81 reflects the circular fringe information into a line in the plane P on which the detector 87 is located. As seen in FIG. 7, the incoming angle θi is reflected onto the plane P at the reflection angle θr, thereby producing a one-to-one mapping with information at radius y of the circle entering the cone being detected at a distance x from the cone apex V. Thus, incoming circular fringe information comprising alternating bright and dark regions are detected as alternating bright and dark regions along the linear array detector 87.
In addition to detecting the presence and size of water droplets, it is desirable to distinguish the cloud phase (water vs. ice) because some engine flameout incidents have been attributed to ingestion of airborne ice crystals by aircraft engines. US Patent Application No. 2010/0110431 to Ray describes a device for optically detecting and distinguishing airborne liquid water droplets and ice crystals includes an illumination portion and a detection portion. The illumination portion outputs a circularly polarized illuminating beam. The detection portion receives circularly polarized backscattered light from moisture in the cloud, in response to the illuminating beam. The circularly polarized backscattered light is passed through a circular polarizer to convert it into linearly polarized backscattered light, which is split into two components. Each of the two components is optionally subject to further linear polarization to filter out any leakage-type orthogonal polarization. The two components are then optically detected and the resulting detection signals are used to calculate one or more parameters reflective of the presence or absence of airborne ice crystals and/or water droplets.